The current cloud property satellite retrieval technique used for ARM and the Clouds and the Earth’s Radiant Energy System (CERES) uses parameterizations of reflectance and effective emittance of cloud layers to simplify the radiative transfer calculations required in determining the optical depth and effective particle size of clouds from space. These retrievals typically assume that the cloud in a given pixel is a single-layered plane-parallel cloud. Because overlapped clouds are common, this assumption often results in cloud properties that represent some mixture of the multilayered cloud properties and do not provide an accurate assessment of actual cloud properties in the scene. Various methods using combinations of infrared and solar channels and solar and microwave data are being used to detect multi-layered clouds. Once detected, it is necessary to unscramble the properties for each of the cloud layers. If it is assumed that the properties of the lower layer cloud are known, it should be possible to derive the properties of the upper-level cloud. This paper investigates the application of the parameterizations used for the CERES and ARM analyses to obtain the upper-level cloud optical depth in multilevel cloud situations. Detailed radiative transfer calculations are used to test the sensitivity and capability of using the visible reflectance parameterization to derive the optical depth of the upper level cloud given the optical depth and particle size of the underlying cloud and the surface albedo. This study will provide guidance as to the use of the visible channel rather than multispectral IR channels to determine the optical depth of the upper cloud layer. The investigation also examines the error introduced in particle size retrieval when the upper level cloud is neglected. Methods for determining the lower-layer cloud proeprties are also discussed.